Probiotics in juvenile idiopathic arthritis: no clinical benefit, and a signal pointing the wrong way

Design

multicentre, randomised, double-blind, placebo-controlled trial

Population

44 children with oligoarticular or RF-negative polyarticular juvenile idiopathic arthritis on standard therapy

Primary outcome

ACR Pedi 30 response at 3 months

Effect

47% with VSL#3 versus 63% with placebo (p=0.33); worst-case analysis 36% versus 68% (p=0.03); faecal TLR4 agonist activity increased

PERMAJI randomised 44 children with oligoarticular or rheumatoid-factor-negative polyarticular juvenile idiopathic arthritis to VSL#3 or placebo for three months alongside standard therapy, double-blind, across multiple French centres. Stool and serum were sampled at baseline and three months for microbiota composition, faecal innate immune agonists, intestinal permeability and systemic cytokines.

ACR Pedi 30 response at three months was 47% with the probiotic and 63% with placebo (p=0.33). Under conservative worst-case handling of missing data, that became 36% versus 68% (p=0.03) - numerically favouring placebo. Faecal Toll-like receptor 4 agonist activity rose significantly with supplementation, while microbiota diversity, intestinal permeability and systemic cytokines were unchanged.

Forty-four children is small, and this is not evidence that probiotics harm children with arthritis. What it does do is remove the assumption that they are a free addition. A supplement that measurably increases exposure to pro-inflammatory microbial stimuli, in a disease driven by immune dysregulation, is not obviously neutral, and parents who ask about probiotics for arthritis can now be given an actual trial result rather than a shrug.

• Do not recommend probiotics as an adjunct in juvenile idiopathic arthritis
• Ask what supplements the family is already giving - probiotics are rarely volunteered as medication
• Answer the parental question with this trial rather than with general reassurance
• Keep the conversation on the treatments with evidence: methotrexate, biologics, and intra-articular steroid
• Note the trial was small - the honest statement is no benefit shown, not harm proven

Why it matters

It converts probiotics for childhood arthritis from an untested but harmless-sounding addition into an intervention with a negative trial behind it.

Don't overread it

Forty-four children: this shows no benefit, and it does not show harm.

The statistics, in plain English

The headline comparison, 47% versus 63% at p=0.33, is a null result in a trial far too small to detect anything but a large effect - 44 children cannot exclude a moderate benefit. The worst-case missing-data analysis reaching p=0.03 in favour of placebo should not be read as showing harm: worst-case imputation is deliberately pessimistic for the treatment arm, and it is a sensitivity analysis, not the primary endpoint. The TLR4 finding is mechanistic and measured in stool, with no clinical consequence demonstrated.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.